The calculation of portfolio unexpected loss in credit and operational risk

Abstract

In effective risk management, it is of importance to take account of risk aspects from both a credit and operational risk viewpoint. This paper attempts to address both of these areas. The essential tool is the calculation of joint default probabilities starting with an asset correlation and a bivariate normal joint distribution assumption, for log returns of asset values in the case of credit risk, and, an event correlation and a bivariate normal joint distribution assumption for performance scores in the case of operational risk. From this it is shown how to calculate default correlations between counterparties or operational risks and hence a method for assessing unexpected loss for portfolios. The method of moments may then be used to construct a portfolio loss distribution and an appropriate extremal loss measure such as 99.9 per cent value at risk (VaR) or tail VaR.